In the era of personalized education, the provision of comprehensible explanations for learning recommendations is of great value to enhance the learner's understanding and engagement with the recommended learning content. Large language models (LLMs) and generative AI have recently opened new doors for generating human-like explanations, for and along learning recommendations. However, their precision is still far away from acceptable in a sensitive field like education. To harness the abilities of LLMs, while still ensuring a high level of precision towards the intent of the learners, this paper proposes an approach to utilize knowledge graphs (KG) as a source of factual context for LLM prompts, reducing the risk of model hallucinations, and safeguarding against wrong or imprecise information, while maintaining an application-intended learning context. We utilize the semantic relations in the knowledge graph to offer curated knowledge about learning recommendations. With domain-experts in the loop, we design the explanation as a textual template, which is filled and completed by the LLM. Domain experts were integrated in the prompt engineering phase as part of a study, to ensure that explanations include information that is relevant to the learner. We evaluate our approach quantitatively using Rouge-N and Rouge-L measures, as well as qualitatively with experts and learners. Our results show an enhanced recall and precision of the generated explanations compared to those generated solely by the GPT model, with a greatly reduced risk of generating imprecise information in the final learning explanation.

In today's e-Business environment, ERP, CRM, collaboration tools, and networked sensors may be characterized as data generators resources. Business Intelligence (BI) is a term that incorporates a range of analytical and decision support applications in business including data mining, decision support systems, knowledge management systems, and online analytical processing; processing data within these systems produce new data that are characterized to grow rapidly causing limitation problem of data management if handled by a Relational Database Management System (RDBMS) or statistical tools. Collectively these structured and unstructured data are referred to as Big Data. Successful and efficient handling of Big Data requires deployment of specific IT infrastructure components as well as adopting an emerging service model. In this research we introduce a conceptual model that abstracts the processing scheme of big data processing lifecycle. The model addresses the main phases of the lifecycle: data acquisition, data serialization, data aggregation, data analysis, data mining, knowledge representation, and information dissemination. The model is driven by projecting Service Oriented Architecture attributes to the building block of the lifecycle and adhering to the Lifecycle Modeling Language specification.

Despite their remarkable success and wide use in many applications, large language models (LLMs) are not free from intrinsic vulnerabilities (e.g. prompt injection). They may also suffer from hallucinations and drop in performance due to lack of up-to-date knowledge. Retrieval-Augmented Generation (RAG) is currently one of the most promising techniques to mitigate such issues. In short, a RAG augments each prompt using a relevant context from an external knowledge database. Usually, the context is composed of texts that are the most similar to the request. While reducing hallucinations, RAG augments at the same time the attack surface of the whole system. Indeed, an attacker may poison the knowledge database by injecting bad or misleading information. In this paper, we introduce HOPRAG, a subtle, but very efficient, poisoning technique that consists in adding a suffix (or prefix) of only few tokens (sub-words) to any given text to raise (or decrease) its similarity with a prompt and therefore be used (or avoid being used) as context by RAG to answer. Our results show that with only three injected tokens, we manage to perform a successful attack.

Objective: Traditional knowledge-based and machine learning diagnostic decision support systems have benefited from integrating the medical domain knowledge encoded in the Unified Medical Language System (UMLS). The emergence of Large Language Models (LLMs) to supplant traditional systems poses questions of the quality and extent of the medical knowledge in the models' internal knowledge representations and the need for external knowledge sources. The objective of this study is three-fold: to probe the diagnosis-related medical knowledge of popular LLMs, to examine the benefit of providing the UMLS knowledge to LLMs (grounding the diagnosis predictions), and to evaluate the correlations between human judgments and the UMLS-based metrics for generations by LLMs. Methods: We evaluated diagnoses generated by LLMs from consumer health questions and daily care notes in the electronic health records using the ConsumerQA and Problem Summarization datasets. Probing LLMs for the UMLS knowledge was performed by prompting the LLM to complete the diagnosis-related UMLS knowledge paths. Grounding the predictions was examined in an approach that integrated the UMLS graph paths and clinical notes in prompting the LLMs. The results were compared to prompting without the UMLS paths. The final experiments examined the alignment of different evaluation metrics, UMLS-based and non-UMLS, with human expert evaluation. Results: In probing the UMLS knowledge, GPT-3.5 significantly outperformed Llama2 and a simple baseline yielding an F1 score of 10.9% in completing one-hop UMLS paths for a given concept. Grounding diagnosis predictions with the UMLS paths improved the results for both models on both tasks, with the highest improvement (4%) in SapBERT score. There was a weak correlation between the widely used evaluation metrics (ROUGE and SapBERT) and human judgments. Conclusion: We found that while popular LLMs contain some medical knowledge in their internal representations, augmentation with the UMLS knowledge provides performance gains around diagnosis generation. The UMLS needs to be tailored for the task to improve the LLMs predictions. Finding evaluation metrics that are aligned with human judgments better than the traditional ROUGE and BERT-based scores remains an open research question.

Prior work on Data-To-Text Generation, the task of converting knowledge graph (KG) triples into natural text, focused on domain-specific benchmark datasets. In this paper, however, we verbalize the entire English Wiki-data KG, and discuss the unique challenges associated with a broad, open-domain, large-scale verbalization. We further show that verbalizing a comprehensive, encyclopedic KG like Wiki-data can be used to integrate structured KGs and natural language corpora. In contrast to the many architectures that have been developed to integrate these two sources, our approach converts the KG into natural text, allowing it to be seamlessly integrated into existing language models. It carries the further advantages of improved factual accuracy and reduced toxicity in the resulting language model. We evaluate this approach by augmenting the retrieval corpus in a retrieval language model and showing significant improvements on the knowledge intensive tasks of open domain QA and the LAMA knowledge probe.

This paper introduces a novel method for determining the best room to place an object in, for embodied scene rearrangement. While state-of-the-art approaches rely on large language models (LLMs) or reinforcement learned (RL) policies for this task, our approach, CLIPGraphs, efficiently combines commonsense domain knowledge, data-driven methods, and recent advances in multimodal learning. Specifically, it (a) encodes a knowledge graph of prior human preferences about the room location of different objects in home environments, (b) incorporates vision-language features to support multimodal queries based on images or text, and (c) uses a graph network to learn object-room affinities based on embeddings of the prior knowledge and the vision-language features. We demonstrate that our approach provides better estimates of the most appropriate location of objects from a benchmark set of object categories in comparison with state-of-the-art baselines. (1)

Linked knowledge graphs build the backbone of many data-driven applications such as search engines, conversational agents and e-commerce solutions. Declarative link discovery frameworks use complex link specifications to express the conditions under which a link between two resources can be deemed to exist. However, understanding such complex link specifications is a challenging task for non-expert users of link discovery frameworks. In this paper, we address this drawback by devising NMV-LS, a language model-based verbalization approach for translating complex link specifications into natural language. NMV-LS relies on the results of rule-based link specification verbalization to apply continuous training on T5, a large language model based on the Transformer architecture. We evaluated NMV-LS on English and German datasets using well-known machine translation metrics such as BLUE, METEOR, ChrF++ and TER. Our results suggest that our approach achieves a verbalization performance close to that of humans and outperforms state of the art approaches. Our source code and datasets are publicly available at https://github.com/dice-group/NMV-LS. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Current language models have a significant limitation in the ability to encode and decode factual knowledge. This is mainly because they acquire such knowledge from statistical co-occurrences although most of the knowledge words are rarely observed. In this paper, we propose a Neural Knowledge Language Model (NKLM) which combines symbolic knowledge provided by the knowledge graph with the RNN language model. By predicting whether the word to generate has an underlying fact or not, the model can generate such knowledge-related words by copying from the description of the predicted fact. In experiments, we show that the NKLM significantly improves the performance while generating a much smaller number of unknown words.

Research publications are the primary vehicle for sharing scientific progress in the form of new discoveries, methods, techniques, and insights. Unfortunately, the lack of a large-scale, comprehensive, and easy-to-use resource capturing the myriad relationships between publications, their authors, and venues presents a barrier to applications for gaining a deeper understanding of science. In this paper, we present PubGraph, a new resource for studying scientific progress that takes the form of a large-scale knowledge graph (KG) with more than 385M entities, 13B main edges, and 1.5B qualifier edges. PubGraph is comprehensive and unifies data from various sources, including Wikidata, OpenAlex, and Semantic Scholar, using the Wikidata ontology. Beyond the metadata available from these sources, PubGraph includes outputs from auxiliary community detection algorithms and large language models. To further support studies on reasoning over scientific networks, we create several large-scale benchmarks extracted from PubGraph for the core task of knowledge graph completion (KGC). These benchmarks present many challenges for knowledge graph embedding models, including an adversarial community-based KGC evaluation setting, zero-shot inductive learning, and large-scale learning. All of the aforementioned resources are accessible at https://pubgraph.isi.edu/ and released under the CC-BY-SA license. We plan to update PubGraph quarterly to accommodate the release of new publications.

For domain-specific speech recognition tasks, it is best if the statistical language model component is trained with text data that is content-wise and style-wise similar to the targeted domain for which the application is built. For state-of-the-art language modeling techniques that can be used in real-time within speech recognition engines during first-pass decoding (e.g., N-gram models), the above constraints have to be fulfilled in the training data. However collecting such data, even through crowd sourcing, is expensive and time consuming, and can still be not representative of how a much larger user population would interact with the recognition system. In this paper, we address this problem by employing several semantic web sources that already contain the domain-specific knowledge, such as query click logs and knowledge graphs. We build statistical language models that meet the requirements listed above for domain-specific recognition tasks where natural language is used and the user queries are about name entities in a specific domain. As a case study, in the movies domain where users' voice queries are movie related, compared to a generic web language model, a language model trained with the above resources not only yields significant perplexity and word-errorrate improvements, but also presents an approach where such language models can be rapidly developed for other domains. Copyright © 2014 ISCA.

Task-oriented dialogue generation is challenging since the underlying knowledge is often dynamic and effectively incorporating knowledge into the learning process is hard. It is particularly challenging to generate both humanlike and informative responses in this setting. Recent research primarily focused on various knowledge distillation methods where the underlying relationship between the facts in a knowledge base is not effectively captured. In this paper, we go one step further and demonstrate how the structural information of a knowledge graph can improve the system's inference capabilities. Specifically, we propose DialoKG, a novel task-oriented dialogue system that effectively incorporates knowledge into a language model. Our proposed system views relational knowledge as a knowledge graph and introduces (1) a structure-aware knowledge embedding technique, and (2) a knowledge graph-weighted attention masking strategy to facilitate the system selecting relevant information during the dialogue generation. An empirical evaluation demonstrates the effectiveness of DialoKG over state-of-theart methods on several standard benchmark datasets. © Findings of the Association for Computational Linguistics: NAACL 2022 - Findings.

The recent advance in neural network architecture and training algorithms has shown the effectiveness of representation learning. The neural-network-based models generate better representation than the traditional ones. They have the ability to automatically learn the distributed representation for sentences and documents. To this end, we proposed a novel model that addresses several issues that are not adequately modeled by the previously proposed models, such as the memory problem and incorporating the knowledge of document structure. Our model uses a hierarchical structured self-attention mechanism to create the sentence and document embeddings. This architecture mirrors the hierarchical structure of the document and in turn enables us to obtain better feature representation. The attention mechanism provides extra source of information to guide the summary extraction. The new model treated the summarization task as a classification problem in which the model computes the respective probabilities of sentence-summary membership. The model predictions are broken up by several features such as information content, salience, novelty, and positional representation. The proposed model was evaluated on two well-known datasets, the CNN/Daily Mail and DUC 2002. The experimental results show that our model outperforms the current extractive state of the art by a considerable margin.

Embedding based Knowledge Graph (KG) Completion has gained much attention over the past few years. Most of the current algorithms consider a KG as a multidirectional labeled graph and lack the ability to capture the semantics underlying the schematic information. In a separate development, a vast amount of information has been captured within the Large Language Models (LLMs) which has revolutionized the field of Artificial Intelligence. KGs could benefit from these LLMs and vice versa. This vision paper discusses the existing algorithms for KG completion based on the variations for generating KG embeddings. It starts with discussing various KG completion algorithms such as transductive and inductive link prediction and entity type prediction algorithms. It then moves on to the algorithms utilizing type information within the KGs, LLMs, and finally to algorithms capturing the semantics represented in different description logic axioms. We conclude the paper with a critical reflection on the current state of work in the community and give recommendations for future directions.

Massive Open Online Courses (MOOCs) have gained popularity in the technology-enhanced learning (TEL) domain. To enhance the learning experience in MOOCs, educational recommender systems (ERSs) can play a crucial role by suggesting courses or learning materials that align with students' knowledge states. Thereby, understanding a student's learning needs and predicting knowledge concepts that the student might be interested in are important to provide effective recommendations. Inspired by the superior ability of knowledge graphs (KGs) in modeling the heterogeneous data in MOOCs and Graph Neural Networks (GNNs) in learning on graph-structured data, few works focusing on GNN-based recommendation of knowledge concepts in MOOCs have emerged recently. However, existing approaches in this domain have limitations mainly related to complexity, semantics, and transparency. To address these limitations, in this paper we propose ConceptGCN, an end-to-end framework that combines KGs, Graph Convolutional Networks (GCNs), and pre-trained transformer language model encoders (SBERT) to provide personalized and transparent recommendations of knowledge concepts in the MOOC platform [Blinded tool]. We conducted extensive offline experiments and an online user study (N=31), demonstrating the benefits of the ConceptGCN-based recommendation approach, in terms of several important user-centric aspects including accuracy, novelty, diversity, usefulness, overall satisfaction, use intentions, and reading intention. In particular, our results indicate that, if SBERT is used for the initial embeddings of items in the KG, a self-connection operation and a semantic similarity-based score function in the aggregation operation of GCN are not necessarily needed. © 2023 The Author(s)

An exciting frontier in natural language understanding (NLU) and generation (NLG) calls for (vision-and-) language models that can efficiently access external structured knowledge repositories. However, many existing knowledge bases only cover limited domains, or suffer from noisy data, and most of all are typically hard to integrate into neural language pipelines. To fill this gap, we release VisualSem: a high-quality knowledge graph (KG) which includes nodes with multilingual glosses, multiple illustrative images, and visually relevant relations. We also release a neural multi-modal retrieval model that can use images or sentences as inputs and retrieves entities in the KG. This multi-modal retrieval model can be integrated into any (neural network) model pipeline. We encourage the research community to use VisualSem for data augmentation and/or as a source of grounding, among other possible uses. VisualSem as well as the multi-modal retrieval models are publicly available and can be downloaded in this URL: https://github.com/iacercalixto/visualsem. © 2021 Association for Computational Linguistics.

With the continuous increase in reported Common Vulnerabilities and Exposures (CVEs), security teams are overwhelmed by vast amounts of data, which are often analyzed manually, leading to a slow and inefficient process. To address cybersecurity threats effectively, it is essential to establish connections across multiple security entity databases, including CVEs, Common Weakness Enumeration (CWEs), and Common Attack Pattern Enumeration and Classification (CAPECs). In this study, we introduce a new approach that leverages the RotatE [4] knowledge graph embedding model, initialized with embeddings from Ada language model developed by OpenAI [3]. Additionally, we extend this approach by initializing the embeddings for the relations. © 2023 Owner/Author(s).

We describe a method, based on Jennifer Nado's definition of classification procedures as targets of conceptual engineering, that implements such procedures using a large language model. We then apply this method using data from the Wikidata knowledge graph to evaluate concept definitions from two paradigmatic conceptual engineering projects: the International Astronomical Union's redefinition of PLANET and Haslanger's ameliorative analysis of WOMAN. We discuss implications of this work for the theory and practice of conceptual engineering. The code and data can be found on GitHub.

While recent vision-and-language models (VLMs) like CLIP are a powerful tool for analyzing text and images in a shared semantic space, they do not explicitly model the hierarchical nature of the set of texts which may describe an image. Conversely, existing multimodal hierarchical representation learning methods require costly training from scratch, failing to leverage the knowledge encoded by state-of-the-art multimodal foundation models. In this work, we study the knowledge of existing foundation models, finding that they exhibit emergent understanding of visual-semantic hierarchies despite not being directly trained for this purpose. We propose the Radial Embedding (RE) framework for probing and optimizing hierarchical understanding, and contribute the HierarCaps dataset, a benchmark facilitating the study of hierarchical knowledge in image–text representations, constructed automatically via large language models. Our results show that foundation VLMs exhibit zero-shot hierarchical understanding, surpassing the performance of prior models explicitly designed for this purpose. Furthermore, we show that foundation models may be better aligned to hierarchical reasoning via a text-only fine-tuning phase, while retaining pretraining knowledge.

An increasing number of security incidents in cyber-physical systems (CPSs) arise from the exploitation of cyber and physical components of such systems. Knowledge about how such incidents arose is rarely captured and used systematically to enhance security and support future incident investigations. In this paper, we propose an approach to represent and share incidents knowledge. Our approach captures incident patterns – common aspects of incidents occurring in different CPSs. Our approach then allows incident patterns to be instantiated for different systems to assess if and how such patterns can manifest again. To support our approach, we provide two meta-models that represent, respectively, incident patterns and the cyber-physical systems themselves. The incident meta-model captures the characteristics of incidents, such as assets and activities. The system meta-model captures cyber and physical components and their interactions, which may be exploited during an incident. We demonstrate the feasibility of our approach in the application domain of smart buildings, by tailoring the system meta-model to represent components and interactions in this domain.

The exponential growth of scientific literature has resulted in information overload, challenging researchers to effectively synthesize relevant publications. This paper explores the integration of traditional reference management software with advanced computational techniques, including Large Language Models and Retrieval-Augmented Generation. We introduce PyZoBot, an AI-driven platform developed in Python, incorporating Zoteros reference management with OpenAIs sophisticated LLMs. PyZoBot streamlines knowledge extraction and synthesis from extensive human-curated scientific literature databases. It demonstrates proficiency in handling complex natural language queries, integrating data from multiple sources, and meticulously presenting references to uphold research integrity and facilitate further exploration. By leveraging LLMs, RAG, and human expertise through a curated library, PyZoBot offers an effective solution to manage information overload and keep pace with rapid scientific advancements. The development of such AI-enhanced tools promises significant improvements in research efficiency and effectiveness across various disciplines.

Explanations are pervasive in our lives. Mostly, they occur in dialogical form where an explainer discusses a concept or phenomenon of interest with an explainee. Leaving the explainee with a clear understanding is not straightforward due to the knowledge gap between the two participants. Previous research looked at the interaction of explanation moves, dialogue acts, and topics in successful dialogues with expert explainers. However, daily-life explanations often fail, raising the question of what makes a dialogue successful. In this work, we study explanation dialogues in terms of the interactions between the explainer and explainee and how they correlate with the quality of explanations in terms of a successful understanding on the explainee's side. In particular, we first construct a corpus of 399 dialogues from the Reddit forum Explain Like I am Five and annotate it for interaction flows and explanation quality. We then analyze the interaction flows, comparing them to those appearing in expert dialogues. Finally, we encode the interaction flows using two language models that can handle long inputs, and we provide empirical evidence for the effectiveness boost gained through the encoding in predicting the success of explanation dialogues. © 2024 ELRA Language Resource Association: CC BY-NC 4.0.

Story components, namely, events, time, participants, and their relations are present in narrative texts from different domains such as journalism, medicine, finance, and law. The automatic extraction of narrative elements encompasses several NLP tasks such as Named Entity Recognition, Semantic Role Labeling, Event Extraction, and Temporal Inference. The text2story Python, an easy-to-use modular library, supports the narrative extraction and visualization pipeline. The package contains an array of narrative extraction tools that can be used separately or in sequence. With this toolkit, end users can process free text in English or Portuguese and obtain formal representations, like standard annotation files or a formal logical representation. The toolkit also enables narrative visualization as Message Sequence Charts (MSC), Knowledge Graphs, and Bubble Diagrams, making it useful to visualize and transform human-annotated narratives. The package combines the use of off-the-shelf and custom tools and is easily patched (replacing existing components) and extended (e.g. with new visualizations). It includes an experimental module for narrative element effectiveness assessment and being is therefore also a valuable asset for researchers developing solutions for narrative extraction. To evaluate the baseline components, we present some results of the main annotators embedded in our package for datasets in English and Portuguese. We also compare the results with the extraction of narrative elements by GPT-3, a robust LLM model. © 2024 ELRA Language Resource Association: CC BY-NC 4.0.

The knowledge graph is a critical resource for medical intelligence. The general medical knowledge graph tries to include all diseases and contains much medical knowledge. However, it is challenging to review all the triples manually. Therefore the quality of the knowledge graph can not support intelligence medical applications. Breast cancer is one of the highest incidences of cancer at present. It is urgent to improve the efficiency of breast cancer diagnosis and treatment through artificial intelligence technology and improve the postoperative health status of breast cancer patients. This paper proposes a framework to construct a breast cancer knowledge graph from heterogeneous data resources in response to this demand. Specifically, this paper extracts knowledge triple from clinical guidelines, medical encyclopedias and electronic medical records. Furthermore, the triples from different data resources are fused to build a breast cancer knowledge graph (BCKG). Experimental results demonstrate that BCKG can support knowledge-based question answering, breast cancer postoperative follow-up and healthcare, and improve the quality and efficiency of breast cancer diagnosis, treatment and management.

Building a knowledge graph is a time-consuming and costly process which often applies complex natural language processing (NLP) methods for extracting knowledge graph triples from text corpora. Pre-trained large Language Models (PLM) have emerged as a crucial type of approach that provides readily available knowledge for a range of AI applications. However, it is unclear whether it is feasible to construct domain-specific knowledge graphs from PLMs. Motivated by the capacity of knowledge graphs to accelerate data-driven materials discovery, we explored a set of state-of-the-art pre-trained general-purpose and domain-specific language models to extract knowledge triples for metal-organic frameworks (MOFs). We created a knowledge graph benchmark with 7 relations for 1248 published MOF synonyms. Our experimental results showed that domain-specific PLMs consistently outperformed the general-purpose PLMs for predicting MOF related triples. The overall benchmarking results, however, show that using the present PLMs to create domain-specific knowledge graphs is still far from being practical, motivating the need to develop more capable and knowledgeable pre-trained language models for particular applications in materials science. © 2022 IEEE.

We present a comprehensive benchmark dataset for Knowledge Graph Question Answering in Materials Science (KGQA4MAT), with a focus on metal-organic frameworks (MOFs). A knowledge graph for metal-organic frameworks (MOF-KG) has been constructed by integrating structured databases and knowledge extracted from the literature. To enhance MOF-KG accessibility for domain experts, we aim to develop a natural language interface for querying the knowledge graph. We have developed a benchmark comprised of 161 complex questions involving comparison, aggregation, and complicated graph structures. Each question is rephrased in three additional variations, resulting in 644 questions and 161 KG queries. To evaluate the benchmark, we have developed a systematic approach for utilizing the LLM, ChatGPT, to translate natural language questions into formal KG queries. We also apply the approach to the well-known QALD-9 dataset, demonstrating ChatGPT's potential in addressing KGQA issues for different platforms and query languages. The benchmark and the proposed approach aim to stimulate further research and development of user-friendly and efficient interfaces for querying domain-specific materials science knowledge graphs, thereby accelerating the discovery of novel materials.

Link prediction and entity resolution play pivotal roles in uncovering hidden relationships within networks and ensuring data quality in the era of heterogeneous data integration. This paper explores the utilization of large language models to enhance link prediction, particularly through knowledge graphs derived from transdisciplinary literature. Investigating zero-shot entity resolution techniques, we examine the impact of ontology-based and large language model approaches on the stability of link prediction results. Through a case study focusing on chronic lower back pain research, we analyze workflow decisions and their influence on prediction outcomes. Our research underscores the importance of robust methodologies in improving predictive accuracy and data integration across diverse domains.

This paper reports an evaluation study for the specification of an average sized expert system for configuring elevator systems using the language KARL (Knowledge Acquisition and Representation Language). Two results have been gained in this study: (i) a formal model of the used problem-solving method (PSM) Propose-and-Revise has been developed and (ii) the adequacy of the language KARL for specifying such systems has been evaluated. KARL is based on a strong conceptual model: the KARL model of expertise, which represents different aspects of the model at different layers. It clearly separates domain specific knowledge fim probleni-solving-specific knowledge which allows to reuse both parts independenty from the other. KARL provides language primitives on a high level of abstraction, independent of implementation issues. KARL is a formal language which allows to represent knowledge unambiguously. KARL is an executable language which allows to validate the resulting model by testing and debugging. It turned out that KARL is well-suited for such specification issues. It also turned out that due to a flexible connection between domain knowledge and problem solving knowledge provided by KARL both different kinds of knowledge may be specified nearly independenty from the other which supports their reuse. This study gave us various insights into the adequacy of the language KARL for representing the knowledge on an abstract level. In spite of the encouraging results we gained this study also revealed some deficiencies of the language KARL which are currently eliminated for a future version of KARL.

Advancements in the capabilities of Large Language Models (LLMs) have created a promising foundation for developing autonomous agents. With the right tools, these agents could learn to solve tasks in new environments by accumulating and updating their knowledge. Current LLM-based agents process past experiences using a full history of observations, summarization, retrieval augmentation. However, these unstructured memory representations do not facilitate the reasoning and planning essential for complex decision-making. In our study, we introduce AriGraph, a novel method wherein the agent constructs and updates a memory graph that integrates semantic and episodic memories while exploring the environment. We demonstrate that our Ariadne LLM agent, consisting of the proposed memory architecture augmented with planning and decision-making, effectively handles complex tasks within interactive text game environments difficult even for human players. Results show that our approach markedly outperforms other established memory methods and strong RL baselines in a range of problems of varying complexity. Additionally, AriGraph demonstrates competitive performance compared to dedicated knowledge graph-based methods in static multi-hop question-answering.

Relationship extraction from unstructured data remains one of the most challenging tasks in the field of Natural Language Processing (NLP). The complexity of relationship extraction arises from the need to comprehend the underlying semantics, syntactic structures, and contextual dependencies within the text. Unstructured data poses challenges with diverse linguistic patterns, implicit relationships, contextual nuances, complicating accurate relationship identification and extraction. The emergence of Large Language Models (LLMs), such as GPT (Generative Pre-trained Transformer), has indeed marked a significant advancement in the field of NLP.In this work, we assess and evaluate the effectiveness of LLMs in relationship extraction in the Holocaust testimonies within the context of the Historical realm. By delving into this domainspecific context, we aim to gain deeper insights into the performance and capabilities of LLMs in accurately capturing and extracting relationships within the Holocaust domain by developing a novel knowledge graph to visualise the relationships of the Holocaust. To the best of our knowledge, there is no existing study which discusses relationship extraction in Holocaust testimonies. The majority of current approaches for Information Extraction (IE) in historic documents are either manual or Optical Character Recognition (OCR) based. Moreover, in this study, we found that the Subject-Object-Verb extraction using GPT3-based relations produced more meaningful results compared to the Semantic Role labelingbased triple extraction. © 2023 Incoma Ltd. All rights reserved.

The Museum of the Person (Museu da Pessoa, MP) is a virtual museum aimed at exhibiting life stories of common people. Its assets are composed of several interviews involving people whose stories we want to perpetuate. So the museum holds an heterogeneous collection of XML (eXtensible Markup Language) documents that constitute the working repository. The main idea is to extract automatically the information included in the repository in order to build the web pages that realize the museum's exhibition rooms. This project started by creating a specific ontology (OntoMP) for the knowledge repository of MP. That ontology is intended to allow a conceptual navigation over the available information. We will adopt the standard for museum ontologies CIDOC-CRM (CIDOC Conceptual Reference Model) refined with FOAF to represent OntoMP. The objective of this paper is to discuss different architectural approaches to build a system that will create the virtual rooms from the XML repository to enable visitors to lookup individual life stories and also intercross information among them. The first architecture is based on a TripleStore and uses SPARQL (SPARQL Protocol and RDF Query Language) technology to extract the information, while the second proposal is based on a Relational Database and uses CaVa Generator to query the repository and build the exhibition spaces.

During their pre-flight briefings, aircraft pilots must analyze a long list of NOTAMs (NOtice To AirMen) indicating potential hazards along the flight route, sometimes up to 100 pages for long-haul flights. NOTAM free-text fields typically have a very special phrasing, with lots of acronyms and domain-specific vocabulary, which makes it differ significantly from standard English. In this paper, we pretrain language models derived from BERT on circa 1 million unlabeled NOTAMs and reuse the learnt representations on three downstream tasks valuable for pilots: criticality prediction, named entity recognition and translation into a structured language called Airlang. This self-supervised approach, where smaller amounts of labeled data are enough for task-specific finetuning, is well suited in the aeronautical context since expert annotations are expensive and time-consuming. We present evaluation scores across the tasks showing a high potential for an operational usability of such models (by pilots, airlines or service providers), which is a first to the best of our knowledge.

The field of natural language understanding has experienced exponential progress in the last few years, with impressive results in several tasks. This success has motivated researchers to study the underlying knowledge encoded by these models. Despite this, attempts to understand their semantic capabilities have not been successful, often leading to non-conclusive, or contradictory conclusions among different works. Via a probing classifier, we extract the underlying knowledge graph of nine of the most influential language models of the last years, including word embeddings, text generators, and context encoders. This probe is based on concept relatedness, grounded on WordNet. Our results reveal that all the models encode this knowledge, but suffer from several inaccuracies. Furthermore, we show that the different architectures and training strategies lead to different model biases. We conduct a systematic evaluation to discover specific factors that explain why some concepts are challenging. We hope our insights will motivate the development of models that capture concepts more precisely.

Clinical problem-solving requires processing of semantic medical knowledge such as illness scripts and numerical medical knowledge of diagnostic tests for evidence-based decision-making. As large language models (LLMs) show promising results in many aspects of language-based clinical practice, their ability to generate non-language evidence-based answers to clinical questions is inherently limited by tokenization. Therefore, we evaluated LLMs' performance on two question types: numeric (correlating findings) and semantic (differentiating entities) while examining differences within and between LLMs in medical aspects and comparing their performance to humans. To generate straightforward multi-choice questions and answers (QAs) based on evidence-based medicine (EBM), we used a comprehensive medical knowledge graph (encompassed data from more than 50,00 peer-reviewed articles) and created the "EBMQA". EBMQA contains 105,000 QAs labeled with medical and non-medical topics and classified into numerical or semantic questions. We benchmarked this dataset using more than 24,500 QAs on two state-of-the-art LLMs: Chat-GPT4 and Claude3-Opus. We evaluated the LLMs accuracy on semantic and numerical question types and according to sub-labeled topics. For validation, six medical experts were tested on 100 numerical EBMQA questions. We found that both LLMs excelled more in semantic than numerical QAs, with Claude3 surpassing GPT4 in numerical QAs. However, both LLMs showed inter and intra gaps in different medical aspects and remained inferior to humans. Thus, their medical advice should be addressed carefully.

While Large Language Models (LLMs) have shown exceptional performance in various tasks, one of their most prominent drawbacks is generating inaccurate or false information with a confident tone. In this paper, we provide evidence that the LLM's internal state can be used to reveal the truthfulness of statements. This includes both statements provided to the LLM, and statements that the LLM itself generates. Our approach is to train a classifier that outputs the probability that a statement is truthful, based on the hidden layer activations of the LLM as it reads or generates the statement. Experiments demonstrate that given a set of test sentences, of which half are true and half false, our trained classifier achieves an average of 71% to 83% accuracy labeling which sentences are true versus false, depending on the LLM base model. Furthermore, we explore the relationship between our classifier's performance and approaches based on the probability assigned to the sentence by the LLM. We show that while LLM-assigned sentence probability is related to sentence truthfulness, this probability is also dependent on sentence length and the frequencies of words in the sentence, resulting in our trained classifier providing a more reliable approach to detecting truthfulness, highlighting its potential to enhance the reliability of LLM-generated content and its practical applicability in real-world scenarios. © 2023 Association for Computational Linguistics.

The development of Activity Based Intelligence (ABI) requires an understanding of individual actors' intents, their interactions with other entities in the environment, and how these interactions facilitate accomplishment of their goals. Statistical modelling alone is insufficient for such analyses, mandating higher-level representations such as ontology to capture important relationships. However, constructing ontologies for ABI, ensuring they remain grounded to real-world entities, and maintaining their applicability to downstream tasks requires substantial hand-tooling by domain experts. In this paper, we propose the use of a Large Language Model (LLM) to bootstrap a grounding for such an ontology. Subsequently, we demonstrate that the experience encoded within the weights of a pre-trained LLM can be used in a zero-shot manner to provide a model of normalcy, enabling ABI analysis at the semantics level, agnostic to the precise coordinate data. This is accomplished through a sequence of two transformations, made upon a kinematic track, toward natural language narratives suitable for LLM input. The first transformation generates an abstraction of the low-level kinematic track, embedding it within a knowledge graph using a domain-specific ABI ontology. Secondly, we employ a template-driven narrative generation process to form natural language descriptions of behavior. Computation of the LLM perplexity score upon these narratives achieves grounding of the ontology. This use does not rely on any prompt engineering. In characterizing the perplexity score for any given track, we observe significant variability given chosen parameters such as sentence verbosity, attribute count, clause ordering, and so on. Consequently, we propose an approach that considers multiple generated narratives for an individual track and the distribution of perplexity scores for downstream applications. We demonstrate the successful application of this methodology against a semantic track association task. Our subsequent analysis establishes how such an approach can be used to augment existing kinematics-based association algorithms.

There has been a surge of interest in utilizing Knowledge Graphs (KGs) for various natural language processing/understanding tasks. The conventional mechanism to retrieve facts in KGs usually involves three steps: entity span detection, entity disambiguation, and relation classification. However, this approach requires additional labels for training each of the three subcomponents in addition to pairs of input texts and facts, and also may accumulate errors propagated from failures in previous steps. To tackle these limitations, we propose a simple knowledge retrieval framework, which directly retrieves facts from the KGs given the input text based on their representational similarities, which we refer to as Direct Fact Retrieval (DiFaR). Specifically, we first embed all facts in KGs onto a dense embedding space by using a language model trained by only pairs of input texts and facts, and then provide the nearest facts in response to the input text. Since the fact, consisting of only two entities and one relation, has little context to encode, we propose to further refine ranks of top-k retrieved facts with a reranker that contextualizes the input text and the fact jointly. We validate our DiFaR framework on multiple fact retrieval tasks, showing that it significantly outperforms relevant baselines that use the three-step approach.

Efficient customer support is the foundation for any service provider trying to improve customer relationships. An important measure of successful support is the mean time to resolve issues. The complexity and large scale of modern cloud environments make it unrealistic to reduce the resolution time without deploying intelligent solutions. The latest also provides an exceptional opportunity to leverage cross-customer product usage data for proactive solutions when the troubles of some users can be analyzed in advance to prevent similar issues of other users. We build a recommender system that matches customer support requests to other resolved support requests or knowledge base articles that contain valuable information for problem remediation. This system can be used by customers or support teams to quickly find problem-resolution tips or detect trending issues to warn vulnerable users. We utilize large language models, fine-tune for better performance, and discuss capabilities and possible improvements. During our research, we highlighted several evaluation metrics such as mean time to resolve issues and the accuracy of recommendations. However, estimating accuracy is challenging due to insufficient datasets with precise and comprehensive recommendations. Despite this, our support managers provided some estimates regarding the remediation durations. Typically, identifying and resolving an issue takes several days or weeks. With appropriate recommendations, this time can be significantly reduced to several hours and, in some simple cases, even lead to self-service capabilities.

Failure mode and effects analysis (FMEA) is a critical tool for mitigating potential failures, particular during ramp-up phases of new products. However, its effectiveness is often limited by the missing reasoning capabilities of the FMEA tools, which are usually tabular structured. Meanwhile, large language models (LLMs) offer novel prospects for fine-tuning on custom datasets for reasoning within FMEA contexts. However, LLMs face challenges in tasks that require factual knowledge, a gap that retrieval-augmented generation (RAG) approaches aim to fill. RAG retrieves information from a non-parametric data store and uses a language model to generate responses. Building on this idea, we propose to advance the non-parametric data store with a knowledge graph (KG). By enhancing the RAG framework with a KG, our objective is to leverage analytical and semantic question-answering capabilities on FMEA data. This paper contributes by presenting a new ontology for FMEA observations, an algorithm for creating vector embeddings from the FMEA KG, and a KG enhanced RAG framework. Our approach is validated through a human study and we measure the performance of the context retrieval recall and precision.

Existing knowledge-grounded conversation systems generate responses typically in a retrieve-then-generate manner. They require a large knowledge base and a strong knowledge retrieval component, which is time- and resource-consuming. In this paper, we address the challenge by leveraging the inherent knowledge encoded in the pre-trained language models (PLMs). We propose Knowledgeable Prefix Tuning (KnowPrefix-Tuning), a two-stage tuning framework, bypassing the retrieval process in a knowledge-grounded conversation system by injecting prior knowledge into the lightweight knowledge prefix. The knowledge prefix is a sequence of continuous knowledge-specific vectors that can be learned during training. In addition, we propose a novel interactive re-parameterization mechanism that allows the prefix to interact fully with the PLM during the optimization of response generation. Experimental results demonstrate that KnowPrefix-Tuning outperforms fine-tuning and other lightweight tuning approaches, and performs comparably with strong retrieval-based baselines while being 3x faster during inference (The code is available at https://github.com/fantast4ever/KnowPrefix-Tuning.)

Existing knowledge-grounded dialogue (KGD) systems access the knowledge from an external knowledge base, then generate the context-coherent response accordingly. However, the knowledge access capability is constrained to the scale of a knowledge base. On the one hand, a small-scale knowledge base makes a model hard to generalize on unseen topics, while the improper shift of topics may induce an unsmooth conversation flow. On the other hand, a large-scale knowledge base requires a strong retrieval component to accurately index the context-relevant knowledge from many plausible candidates, costing significant amounts of time and resources. To address this, we regard the language model as a virtual knowledge base and propose homogenizing internalized knowledge of different language models into hybrid prompts. The hybrid prompts are a set of continuous vectors learned to represent knowledge inherently encoded in different language models. Furthermore, we devise a two-stage knowledge-grounding manner, in which both the knowledge internalized in language models and the knowledge provided by evidence can be jointly optimized to generate a knowledgeable response. We compare our proposed method with two groups of methods, including methods with explicit knowledge retrieval and those with implicit knowledge access. Experimental results on three knowledge-grounded dialogue corpora demonstrate advantages over these competitive methods.

Integrating Large Language Models (LLMs) with logic-based Enterprise Knowledge Graphs (EKGs) and more generally with Knowledge Representation and Reasoning (KRR) approaches is currently at the forefront of research in many data-intensive areas, as language models may complement EKGs and ontological reasoning with flexibility and human orientation. Conversely, EKGs provide transparency and explainability on the conclusions drawn, a typical weak point of LLMs, which operate opaquely. In this demo, we integrate Llama 2 with our reasoning system Vadalog and use it to turn a chase graph, i.e., the trace of an ontological reasoning process, into a human-readable business report. In other words, we show the amazing capabilities of state-of-the-art LLMs in combination with a principled exploitation of the theoretical underpinnings of logic-based reasoning. We walk the audience through a visual environment, unfolding real-world reasoning settings from the Central Bank of Italy. © 2024 Copyright held by the owner/author(s).

In recent times, the demand for transparency and accountability in AI-driven decisions has intensified, particularly in high-stakes domains like finance and bio-medicine. This focus on the provenance of AI-generated conclusions underscores the need for decision-making processes that are not only transparent but also readily interpretable by humans, to built trust of both users and stakeholders. In this context, the integration of state-of-the-art Large Language Models (LLMs) with logic-oriented Enterprise Knowledge Graphs (EKGs) and the broader scope of Knowledge Representation and Reasoning (KRR) methodologies is currently at the cutting edge of industrial and academic research across numerous data-intensive areas. Indeed, such a synergy is paramount as LLMs bring a layer of adaptability and human-centric understanding that complements the structured insights of EKGs. Conversely, the central role of ontological reasoning is to capture the domain knowledge, accurately handling complex tasks over a given realm of interest, and to infuse the process with transparency and a clear provenance-based explanation of the conclusions drawn, addressing the fundamental challenge of LLMs' inherent opacity and fostering trust and accountability in AI applications. In this paper, we propose a novel neuro-symbolic framework that leverages the underpinnings of provenance in ontological reasoning to enhance state-of-the-art LLMs with domain awareness and explainability, enabling them to act as natural language interfaces to EKGs. © Teodoro Baldazzi, Luigi Bellomarini, Stefano Ceri, Andrea Colombo, Andrea Gentili, Emanuel Sallinger, and Paolo Atzeni; licensed under Creative Commons License CC-BY 4.0.

Question answering (QA)-producing correct answers for input questions-is popular, but we test a reverse question answering (RQA) task: given an input answer, generate a question with that answer. Past work tests QA and RQA separately, but we test them jointly, comparing their difficulty, aiding benchmark design, and assessing reasoning consistency. 16 LLMs run QA and RQA with trivia questions/answers, showing: 1) Versus QA, LLMs are much less accurate in RQA for numerical answers, but slightly more accurate in RQA for textual answers; 2) LLMs often answer their own invalid questions from RQA accurately in QA, so RQA errors are not from knowledge gaps alone; 3) RQA errors correlate with question difficulty and inversely correlate with answer frequencies in the Dolma corpus; and 4) LLMs struggle to give valid multi-hop questions. By finding question and answer types yielding RQA errors, we suggest improvements for LLM RQA reasoning.

In this work, we analyse the role of output vocabulary for text-to-text (T2T) models on the task of SPARQL semantic parsing. We perform experiments within the the context of knowledge graph question answering (KGQA), where the task is to convert questions in natural language to the SPARQL query language. We observe that the query vocabulary is distinct from human vocabulary. Language Models (LMs) are pre-dominantly trained for human language tasks, and hence, if the query vocabulary is replaced with a vocabulary more attuned to the LM tokenizer, the performance of models may improve. We carry out carefully selected vocabulary substitutions on the queries and find absolute gains in the range of 17% on the GrailQA dataset. © 2023 Association for Computational Linguistics.

The aim of all Question Answering (QA) systems is to generalize to unseen questions. Current supervised methods are reliant on expensive data annotation. Moreover, such annotations can introduce unintended annotator bias, making systems focus more on the bias than the actual task. This work proposes Knowledge Triplet Learning (KTL), a self-supervised task over knowledge graphs. We propose heuristics to create synthetic graphs for commonsense and scientific knowledge. We propose using KTL to perform zero-shot question answering, and our experiments show considerable improvements over large pre-trained transformer language models.

Automatic term extraction is the process of identifying domain-specific terms in a text using automated algorithms and is a key first step in ontology learning and knowledge graph creation. Large language models have shown good few-shot capabilities, thus, in this paper, we present a study to evaluate the few-shot in-context learning performance of GPT-3.5-Turbo on automatic term extraction. To benchmark the performance we compare the results with fine-tuning of a BERT-sized model. We also carry out experiments with count-based term extractors to assess their applicability to few-shot scenarios. We quantify prompt sensitivity with experiments to analyze the variation in performance of large language models across different prompt templates. Our results show that in-context learning with GPT-3.5-Turbo outperforms the BERT-based model and unsupervised count-based methods in few-shot scenarios.

This paper proposes a chatbot framework that adopts a hybrid model which consists of a knowledge graph and a text similarity model. Based on this chatbot framework, we build HHH, an online question-and-answer (QA) Healthcare Helper system for answering complex medical questions. HHH maintains a knowledge graph constructed from medical data collected from the Internet. HHH also implements a novel text representation and similarity deep learning model, Hierarchical BiLSTM Attention Model (HBAM), to find the most similar question from a large QA dataset. We compare HBAM with other state-of-the-art language models such as bidirectional encoder representation from transformers (BERT) and Manhattan LSTM Model (MaLSTM). We train and test the models with a subset of the Quora duplicate questions dataset in the medical area. The experimental results show that our model is able to achieve a superior performance than these existing methods.

We propose DISC-MedLLM, a comprehensive solution that leverages Large Language Models (LLMs) to provide accurate and truthful medical response in end-to-end conversational healthcare services. To construct high-quality Supervised Fine-Tuning (SFT) datasets, we employ three strategies: utilizing medical knowledge-graphs, reconstructing real-world dialogues, and incorporating human-guided preference rephrasing. These datasets are instrumental in training DISC-MedLLM, surpassing existing medical LLMs in both single-turn and multi-turn consultation scenarios. Extensive experimental results demonstrate the effectiveness of the proposed model in bridging the gap between general language models and real-world medical consultation. Additionally, we release the constructed dataset and model weights to further contribute to research and development. Further details and resources can be found at https://github.com/FudanDISC/DISC-MedLLM

Large language models (LLMs) can generate long-form and coherent text, yet they often hallucinate facts, which undermines their reliability. To mitigate this issue, inference-time methods steer LLM representations toward the “truthful directions” previously learned for truth elicitation. However, applying these truthful directions with the same intensity fails to generalize across different query contexts. We propose LITO, a Learnable Intervention method for Truthfulness Optimization that automatically identifies the optimal intervention intensity tailored to each specific context. LITO explores a sequence of model generations based on increasing levels of intervention intensities. It selects the most accurate response or refuses to answer when the predictions are highly uncertain. Experiments on multiple LLMs and question-answering datasets demonstrate that LITO improves truthfulness while preserving task accuracy. The adaptive nature of LITO counters the limitations of one-size-fits-all intervention methods, maximizing truthfulness by reflecting the model's internal knowledge only when it is confident. Our code is available at https://github.com/launchnlp/LITO. © 2024 Association for Computational Linguistics.

Understanding contextual behavior is very important in order to develop a context-aware retrieval system. This paper discusses the philosophy behind the development of the "evolutionary behavior of textual semantics" (EBOTS) system. The EBOTS system is retrieval oriented knowledge representation and management system. This paper proposes a formal model of correlation that can be combined with traditional local and global weighing schemes. Intuitive contextual behavior is studied as a part of proposed research work. Context retrieval based on semantic knowledge allows abstract queries to be defined, instead of exact word-based queries. The results of the context retrieval for a classic3 and time dataset using the EBOTS system have been discussed in this paper. The paper makes a contribution to the semantic knowledge representation and retrieval algorithms.

Despite their vast capabilities, Large Language Models (LLMs) often struggle with generating reliable outputs, frequently producing high-confidence inaccuracies known as hallucinations. Addressing this challenge, our research introduces InternalInspector, a novel framework designed to enhance confidence estimation in LLMs by leveraging contrastive learning on internal states including attention states, feed-forward states, and activation states of all layers. Unlike existing methods that primarily focus on the final activation state, InternalInspector conducts a comprehensive analysis across all internal states of every layer to accurately identify both correct and incorrect prediction processes. By benchmarking InternalInspector against existing confidence estimation methods across various natural language understanding and generation tasks, including factual question answering, commonsense reasoning, and reading comprehension, InternalInspector achieves significantly higher accuracy in aligning the estimated confidence scores with the correctness of the LLM's predictions and lower calibration error. Furthermore, InternalInspector excels at HaluEval, a hallucination detection benchmark, outperforming other internal-based confidence estimation methods in this task.

The automated construction of knowledge graphs from procedural documents is a challenging research area. Here, the lack of annotated data, as well as raw text repositories describing real-world procedural documents, make it extremely difficult to adopt deep learning approaches. Pre-trained language models have shown promising results concerning the knowledge extraction tasks from the models themselves. Although several works explored this strategy to build knowledge graph, the viability of knowledge base construction by using prompt-based learning strategy from such language models has not yet been investigated deeply. In this work, we present a prompt-based in-context learning strategy to extract, from natural language process descriptions, conceptual information that can be converted into their equivalent knowledge graphs. Such a strategy is performed in a multi-turn dialog fashion. We validate the accuracy of the proposed approach from both quantitative and qualitative perspectives. The results highlight the feasibility of the proposed approach within low-resource scenarios.

The automated construction of process knowledge graphs from process description documents is a challenging research area. Here, the lack of massive annotated data, as well as raw text repositories describing real-world process documents, makes it extremely difficult to adopt deep learning approaches to perform this transformation. Indeed, the main challenge is to extract conceptual elements representing the actual entities or relations of the process model described within its corresponding natural language document. Large Language Models (LLMs) have shown promising results in supporting the extraction of structured knowledge from unstructured texts. Although several works explored this strategy to build or complete knowledge graphs, the exploitation of LLMs toward domain-specific knowledge base construction from scratch has not yet been investigated deeply. Our aim is to exploit the LLM capabilities to extract process knowledge from unseen natural language descriptions. In this work, we present a prompt-based in-context learning strategy to extract, from process descriptions, conceptual information that can be converted into their equivalent knowledge graphs. Such a strategy is performed in a multi-turn dialog fashion. We validate the accuracy of the proposed approach from a quantitative perspective. The results highlight the feasibility of the proposed approach within our low-resource scenarios and open interesting perspectives for future activities.

We present KNOW–the Knowledge Navigator Ontology for the World–the first ontology designed to capture everyday knowledge to augment large language models (LLMs) in real-world generative AI use cases such as personal AI assistants. Our domain is human life, both its everyday concerns and its major milestones. We have limited the initial scope of the modeled concepts to only established human universals: spacetime (places, events) plus social (people, groups, organizations). The inclusion criteria for modeled concepts are pragmatic, beginning with universality and utility. We compare and contrast previous work such as Schema.org and Cyc–as well as attempts at a synthesis of knowledge graphs and language models–noting how LLMs already encode internally much of the commonsense tacit knowledge that took decades to capture in the Cyc project. We also make available code-generated software libraries for the 12 most popular programming languages, enabling the direct use of ontology concepts in software engineering. We emphasize simplicity and developer experience in promoting AI interoperability.

Automated and explainable data interpretation hinges on two critical steps: (i) identifying emerging properties from data and representing them into abstract concepts, and (ii) translating such concepts into natural language. While Large Language Models have recently demonstrated impressive capabilities in generating natural language, their trustworthiness remains difficult to ascertain. The deployment of an explainable pipeline enables its application in high-risk activities, such as decision making. Addressing this demanding requirement is facilitated by the fertile ground of knowledge representation and automated reasoning research. Building upon previous work that explored the first step, we focus on the second step, named Concept2Text. The design of an explainable translation naturally lends itself to a logic-based model, once again highlighting the contribution of declarative programming to achieving explainability in AI. This paper explores a Prolog/CLP-based rewriting system designed to interpret concepts expressed in terms of classes and relations derived from a generic ontology, generating text in natural language. Its key features encompass hierarchical tree rewritings, modular multilingual generation, support for equivalent variants across semantic, grammar, and lexical levels, and a transparent rule-based system. We present the architecture and illustrate a simple working example that allows the generation of hundreds of different and equivalent rewritings relative to the input concept. © 2024 Copyright for this paper by its authors.

In some areas of software engineering research, there are several metamodels claiming to capture the main issues. Though it is profitable to have variety at the beginning of a research field, after some time, the diversity of metamodels becomes an obstacle, for instance to the sharing of results between research groups. To reach consensus and unification of existing metamodels, metamodel-driven software language engineering can be applied. This paper illustrates an application of software language engineering in the agent-oriented software engineering research domain. Here, we introduce a relatively generic agent-oriented metamodel whose suitability for supporting modeling language development is demonstrated by evaluating it with respect to several existing methodology-specific metamodels. First, the metamodel is constructed by a combination of bottom-up and top-down analysis and best practice. The concepts thus obtained and their relationships are then evaluated by mapping to two agent-oriented metamodels: TAO and Islander. We then refine the metamodel by extending the comparisons with the metamodels implicit or explicit within five more extant agent-oriented approaches: Adelfe, PASSI, Gaia, INGENIAS, and Tropos. The resultant FAML metamodel is a potential candidate for future standardization as an important component for engineering an agent modeling language.

Language Models such as BERT (Bidirectional Encoder Representations from Transformers) have grown in popularity due to their ability to be pre-trained and perform robustly on a wide range of Natural Language Processing tasks. Often seen as an evolution over traditional word embedding techniques, they can produce semantic representations of text, useful for tasks such as semantic similarity. However, state-of-the-art models often have high computational requirements and lack global context or domain knowledge which is required for complete language understanding. To address these limitations, we investigate the benefits of knowledge incorporation into the fine-tuning stages of BERT. An existing K-BERT model, which enriches sentences with triplets from a Knowledge Graph, is adapted for the English language and extended to inject contextually relevant information into sentences. As a sideeffect, changes made to K-BERT for accommodating the English language also extend to other word-based languages. Experiments conducted indicate that injected knowledge introduces noise. We see statistically significant improvements for knowledge-driven tasks when this noise is minimised. We show evidence that, given the appropriate task, modest injection with relevant, high-quality knowledge is most performant.

The market debut of ChatGPT gave rise to the development and deployment of various other Large Language Models (LLMs) that achieve state-of-the-art performance across various tasks. The growing popularity of these models has captivated some to attempt to construct or enhance their own LLM. We must be aware of the significant problems that already exist and that we might face along the way. This paper aims to identify and investigate the main challenges in this field, provide existing solutions, and propose novel approaches to mitigate them. A unique Truth-Table proposition for validating synthetic data is presented examining two models, along with a bidirectional knowledge graph-based solution for curing the reverse curse problem, data generation strategies, domain adaptation methods, and the use of a custom dataset to address model hallucinations. The methodology and findings of this study provide valuable insights for users, researchers, and industry experts who are interested in LLMs. It serves as a reference for future research on current models, refining models or developing domain-specific ones. © 2024 IEEE.

Induction-the ability to generalize from existing knowledge-is the cornerstone of intelligence. Cognitive models of human induction are largely limited to toy problems and cannot make quantitative predictions for the thousands of different induction arguments that have been studied by researchers, or to the countless induction arguments that could be encountered in everyday life. Leading large language models (LLMs) go beyond toy problems but fail to mimic observed patterns of human induction. In this article, we combine rich knowledge representations obtained from LLMs with theories of human inductive reasoning developed by cognitive psychologists. We show that this integrative approach can capture several benchmark empirical findings on human induction and generate human-like responses to natural language arguments with thousands of common categories and properties. These findings shed light on the cognitive mechanisms at play in human induction and show how existing theories in psychology and cognitive science can be integrated with new methods in artificial intelligence, to successfully model high-level human cognition.

Large Language Models (LLMs) have demonstrated potential in cybersecurity applications but have also caused lower confidence due to problems like hallucinations and a lack of truthfulness. Existing benchmarks provide general evaluations but do not sufficiently address the practical and applied aspects of LLM performance in cybersecurity-specific tasks. To address this gap, we introduce the SECURE (Security Extraction, Understanding & Reasoning Evaluation), a benchmark designed to assess LLMs performance in realistic cybersecurity scenarios. SECURE includes six datasets focussed on the Industrial Control System sector to evaluate knowledge extraction, understanding, and reasoning based on industry-standard sources. Our study evaluates seven state-of-the-art models on these tasks, providing insights into their strengths and weaknesses in cybersecurity contexts, and offer recommendations for improving LLMs reliability as cyber advisory tools.

Current generative knowledge graph construction approaches usually fail to capture structural knowledge by simply flattening natural language into serialized texts or a specification language. However, large generative language model trained on structured data such as code has demonstrated impressive capability in understanding natural language for structural prediction and reasoning tasks. Intuitively, we address the task of generative knowledge graph constructionwith code languagemodel: given a code-format natural language input, the target is to generate triples which can be represented as code completion tasks. Specifically, we develop schema-aware prompts that effectively utilize the semantic structure within the knowledge graph. As code inherently possesses structure, such as class and function definitions, it serves as a useful model for prior semantic structural knowledge. Furthermore, we employ a rationale-enhanced generation method to boost the performance. Rationales provide intermediate steps, thereby improving knowledge extraction abilities. Experimental results indicate that the proposed approach can obtain better performance on benchmark datasets compared with baselines.

Machine learning is a subfield of artificial intelligence that deals with the exploration and construction of systems that can learn from data. Machine learning trains the computers to manage the critical situations via examining, self-training, inference by observation and previous experience. This paper provides an overview of the development of an efficient classifier that represents the semantics in medical data (Medline) using a Machine Learning (ML) perspective. In recent days people are more concerned about their health and explore ways to identify health related information. But the process of identifying the semantic representation for the medical terms is a difficult task. The main goal of our work was to identify the semantic representation for the medical abstracts in the Medline repository using Machine Learning and Natural Language Processing (NLP).

The entity type information in Knowledge Graphs (KGs) such as DBpedia, Freebase, etc. is often incomplete due to automated generation or human curation. Entity typing is the task of assigning or inferring the semantic type of an entity in a KG. This paper presents GRAND, a novel approach for entity typing leveraging different graph walk strategies in RDF2vec together with textual entity descriptions. RDF2vec first generates graph walks and then uses a language model to obtain embeddings for each node in the graph. This study shows that the walk generation strategy and the embedding model have a significant effect on the performance of the entity typing task. The proposed approach outperforms the baseline approaches on the benchmark datasets DBpedia and FIGER for entity typing in KGs for both fine-grained and coarse-grained classes. The results show that the combination of orderaware RDF2vec variants together with the contextual embeddings of the textual entity descriptions achieve the best results.

Knowledge Graphs (KGs) have become the backbone of various machine learning based applications over the past decade. However, the KGs are often incomplete and inconsistent. Several representation learning based approaches have been introduced to complete the missing information in KGs. Besides, Neural Language Models (NLMs) have gained huge momentum in NLP applications. However, exploiting the contextual NLMs to tackle the Knowledge Graph Completion (KGC) task is still an open research problem. In this paper, a GPT-2 based KGC model is proposed and is evaluated on two benchmark datasets. The initial results obtained from the fine-tuning of the GPT-2 model for triple classification strengthens the importance of usage of NLMs for KGC. Also, the impact of contextual language models for KGC has been discussed. © 2021 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).

The entity type information in Knowledge Graphs (KGs) such as DBpedia, Freebase, etc. is often incomplete due to automated generation. Entity Typing is the task of assigning or inferring the semantic type of an entity in a KG. This paper introduces an approach named Cat2Type which exploits the Wikipedia Categories to predict the missing entity types in a KG. This work extracts information from Wikipedia Category names and the Wikipedia Category graph which are the sources of rich semantic information about the entities. In Cat2Type, the characteristic features of the entities encapsulated in Wikipedia Category names are exploited using Neural Language Models. On the other hand, a Wikipedia Category graph is constructed to capture the connection between the categories. The Node level representations are learned by optimizing the neighbourhood information on the Wikipedia category graph. These representations are then used for entity type prediction via classification. The performance of Cat2Type is assessed on two real-world benchmark datasets DBpedia630k and FIGER. The experiments depict that Cat2Type obtained a significant improvement over state-of-the-art approaches.

The task of Visual Question Answering (VQA) has been studied extensively on general-domain real-world images. Transferring insights from general domain VQA to the art domain (ArtVQA) is non-trivial, as the latter requires models to identify abstract concepts, details of brushstrokes and styles of paintings in the visual data as well as possess background knowledge about art. This is exacerbated by the lack of high-quality datasets. In this work, we shed light on hidden linguistic biases in the AQUA dataset, which is the only publicly available benchmark dataset for ArtVQA. As a result, the majority of questions can be answered without consulting the visual information, making the “V” in ArtVQA rather insignificant. In order to counter this problem, we create a simple, yet practical dataset, ArtQuest, using structured information from the SemArt collection. Our dataset and the pipeline to reproduce our results are publicly available at https://github.com/bletib/artquest.

Large Language Models (LLMs) frequently lack domain-specific knowledge and even fine-tuned models tend to hallucinate. Hence, more reliable models that can include external knowledge are needed. We present a pipeline, 4StepFocus, and specifically a preprocessing step, that can substantially improve the answers of LLMs. This is achieved by providing guided access to external knowledge making use of the model's ability to capture relational context and conduct rudimentary reasoning by themselves. The method narrows down potentially correct answers by triplets-based searches in a semi-structured knowledge base in a direct, traceable fashion, before switching to latent representations for ranking those candidates based on unstructured data. This distinguishes it from related methods that are purely based on latent representations. 4StepFocus consists of the steps: 1) Triplet generation for extraction of relational data by an LLM, 2) substitution of variables in those triplets to narrow down answer candidates employing a knowledge graph, 3) sorting remaining candidates with a vector similarity search involving associated non-structured data, 4) reranking the best candidates by the LLM with background data provided. Experiments on a medical, a product recommendation, and an academic paper search test set demonstrate that this approach is indeed a powerful augmentation. It not only adds relevant traceable background information from information retrieval, but also improves performance considerably in comparison to state-of-the-art methods. This paper presents a novel, largely unexplored direction and therefore provides a wide range of future work opportunities. Used source code is available at https://github.com/kramerlab/4StepFocus.

Pre-trained language models are now ubiquitous in natural language processing, being successfully applied for many different tasks and in several real-world applications. However, even though there is a wealth of high-quality written materials on surgery, and the scientific community has shown a growing interest in the application of natural language processing techniques in surgery, a pre-trained language model specific to the surgical domain is still missing. The creation and public release of such a model would serve numerous useful clinical applications. For example, it could enhance existing surgical knowledge bases employed for task automation, or assist medical students in summarizing complex surgical descriptions. For this reason, in this paper, we introduce SurgicBERTa, a pre-trained language model specific for the English surgical language, i.e., the language used in the surgical domain. SurgicBERTa has been obtained from RoBERTa through continued pre-training with the Masked language modeling objective on 300 k sentences taken from English surgical books and papers, for a total of 7 million words. By publicly releasing SurgicBERTa, we make available a resource built from the content collected in many high-quality surgical books, online textual resources, and academic papers. We performed several assessments in order to evaluate SurgicBERTa, comparing it with the general domain RoBERTa. First, we intrinsically assessed the model in terms of perplexity, accuracy, and evaluation loss resulting from the continual training according to the masked language modeling task. Then, we extrinsically evaluated SurgicBERTa on several downstream tasks, namely (i) procedural sentence detection, (ii) procedural knowledge extraction, (iii) ontological information discovery, and (iv) surgical terminology acquisition. Finally, we conducted some qualitative analysis on SurgicBERTa, showing that it contains a lot of surgical knowledge that could be useful to enrich existing state-of-the-art surgical knowledge bases or to extract surgical knowledge. All the assessments show that SurgicBERTa better deals with surgical language than a general-purpose pre-trained language model such as RoBERTa, and therefore can be effectively exploited in many computer-assisted applications in the surgical domain.

In systems analysis and design it is common to refer to some widely used de-facto industry standards like Unified Modeling Language (UML) and Business Process Model and Notation (BPMN). Albeit the wide adoption of such standard modeling languages, only limited research focuses on the techniques in which these standards are specified and the quality they provide. Most research focuses on case studies of applying standards, ways of extending standards to domain-specific requirements, e.g., by means of profiling, or evaluations of single modeling languages, e.g., using questionnaires or semiotic theories. By contrast, this paper critically reflects on the current state of modeling standards with a focus on their graphical representation (notation). The contribution of this paper is threefold: First, a systematic analysis is performed thereby investigating how different modeling standards specify notational aspects. Second, an evaluation is performed by applying Moody's Physics of Notation theory to the identified standards. Third, based on the findings, recommendations are given to improve modeling standard specifications in the future w.r.t. their notational aspects.

Tackling the need of large organizations for a proactive Asset Information Management (AIM) System, a methodological approach to knowledge management applied to built assets portfolios is proposed. It aims at synergically leveraging Knowledge Graphs (KGs) and Artificial Intelligence (AI) technologies to enable analytics on input data. In the theorized pipeline Large Language Models (LLMs) are meant to be used both in the graph creation phase, extracting data from unstructured sources and organizing them according to domain ontologies, as tested on a use-case sample, and in the knowledge extraction phase via queries. © 2024 European Council on Computing in Construction.

We present the first comprehensive study on automatic knowledge base construction for two prevalent commonsense knowledge graphs: ATOMIC (Sap et al., 2019) and ConceptNet (Speer et al., 2017). Contrary to many conventional 103s that store knowledge with canonical templates, commonsense KBs only store loosely structured open-text descriptions of knowledge. We posit that an important step toward automatic commonsense completion is the development of generative models of commonsense knowledge, and propose COMmonsEnse Transformers (COMET(sic)) that learn to generate rich and diverse conunonsense descriptions in natural language. Despite the challenges of commonsense modeling, our investigation reveals promising results when implicit knowledge from deep pre-trained language models is transferred to generate explicit knowledge in commonsense knowledge graphs. Empirical results demonstrate that COMET is able to generate novel knowledge that humans rate as high quality, with up to 77.5% (ATOMIC) and 91.7% (ConceptNet) precision at top 1, which approaches human performance for these resources. Our findings suggest that using generative commonsense models for automatic commonsense KB completion could soon be a plausible alternative to extractive methods.

The PECO framework is a knowledge representation for formulating clinical questions. Queries are decomposed into four aspects, which are Patient-Problem (P), Exposure (E), Comparison (C) and Outcome (O). However, no test collection is available to evaluate such framework in information retrieval. In this work, we first present the construction of a large test collection extracted from systematic literature reviews. We then describe an analysis of the distribution of PECO elements throughout the relevant documents and propose a language modeling approach that uses these distributions as a weighting strategy. In our experiments carried out on a collection of 1.5 million documents and 423 queries, our method was found to lead to an improvement of 28% in MAP and 50% in P@5, as compared to the state-of-the-art method. © 2010 Association for Computational Linguistics.

The efficiency of information retrieval systems primarily depends on the effective representation of documents during query processing. This representation is mainly constructed from relevant document terms identified and selected during their indexing, which are then used for retrieval. However, when documents contain only a few features, such as in short documents, the resulting representation may be information-poor due to a lack of index terms and their lack of relevance. Although document representation can be enriched using techniques like word embeddings, these techniques require large pre-trained datasets, which are often unavailable in the context of domain-specific short documents. This study investigates a new approach to enrich document representation during indexing using generative AI. In the proposed approach, relevant terms extracted from documents and preprocessed for indexing are enriched with a list of key terms suggested by a large language model (LLM). After conducting a small benchmark of several renowned LLM models for key term suggestions from a set of short texts, the GPT-4o model was chosen to experiment with the proposed indexing approach. The findings of this study yielded notable results, demonstrating that generative AI can efficiently fill the knowledge gap in document representation, regardless of the retrieval technique used.

Large Language Models (LLMs) demonstrate an impressive capacity to recall a vast range of factual knowledge. However, understanding their underlying reasoning and internal mechanisms in exploiting this knowledge remains a key research area. This work unveils the factual information an LLM represents internally for sentence-level claim verification. We propose an end-to-end framework to decode factual knowledge embedded in token representations from a vector space to a set of ground predicates, showing its layer-wise evolution using a dynamic knowledge graph. Our framework employs activation patching, a vector-level technique that alters a token representation during inference, to extract encoded knowledge. Accordingly, we neither rely on training nor external models. Using factual and common-sense claims from two claim verification datasets, we showcase interpretability analyses at local and global levels. The local analysis highlights entity centrality in LLM reasoning, from claim-related information and multi-hop reasoning to representation errors causing erroneous evaluation. On the other hand, the global reveals trends in the underlying evolution, such as word-based knowledge evolving into claim-related facts. By interpreting semantics from LLM latent representations and enabling graph-related analyses, this work enhances the understanding of the factual knowledge resolution process.

For centuries, researchers have sought out ways to connect disparate areas of knowledge. While early scholars (Galileo, da Vinci, etc.) were experts across fields, specialization took hold later. With the advent of Artificial Intelligence, we can now explore relationships across areas (e.g., mechanics-biology) or disparate domains (e.g., failure mechanics-art). To achieve this, we use a fine-tuned large language model (LLM), here for a subset of knowledge in multiscale materials failure. The approach includes the use of a general-purpose LLM to distill question-answer pairs from raw sources followed by LLM fine-tuning. The resulting MechGPT LLM foundation model is used in a series of computational experiments to explore its capacity for knowledge retrieval, various language tasks, hypothesis generation, and connecting knowledge across disparate areas. While the model has some ability to recall knowledge from training, we find that LLMs are particularly useful for extracting structural insights through Ontological Knowledge Graphs. These interpretable graph structures provide explanatory insights, frameworks for new research questions, and visual representations of knowledge that also can be used in retrieval-augmented generation. Three versions of MechGPT are discussed, featuring different sizes from 13 x 109 to 70 x 109 parameters, and reaching context lengths of more than 10,000 tokens. This provides ample capacity for sophisticated retrieval augmented strategies, as well as agent-based modeling where multiple LLMs interact collaboratively and/or adversarially, the incorporation of new data from the literature or web searches, as well as multimodality.

Transformer neural networks show promising capabilities, in particular for uses in materials analysis, design, and manufacturing, including their capacity to work effectively with human language, symbols, code, and numerical data. Here, we explore the use of large language models (LLMs) as a tool that can support engineering analysis of materials, applied to retrieving key information about subject areas, developing research hypotheses, discovery of mechanistic relationships across disparate areas of knowledge, and writing and executing simulation codes for active knowledge generation based on physical ground truths. Moreover, when used as sets of AI agents with specific features, capabilities, and instructions, LLMs can provide powerful problem-solution strategies for applications in analysis and design problems. Our experiments focus on using a fine-tuned model, MechGPT, developed based on training data in the mechanics of materials domain. We first affirm how fine-tuning endows LLMs with a reasonable understanding of subject area knowledge. However, when queried outside the context of learned matter, LLMs can have difficulty recalling correct information and may hallucinate. We show how this can be addressed using retrieval-augmented Ontological Knowledge Graph strategies. The graph-based strategy helps us not only to discern how the model understands what concepts are important but also how they are related, which significantly improves generative performance and also naturally allows for injection of new and augmented data sources into generative AI algorithms. We find that the additional feature of relatedness provides advantages over regular retrieval augmentation approaches and not only improves LLM performance but also provides mechanistic insights for exploration of a material design process. Illustrated for a use case of relating distinct areas of knowledge, here, music and proteins, such strategies can also provide an interpretable graph structure with rich information at the node, edge, and subgraph level that provides specific insights into mechanisms and relationships. We discuss other approaches to improve generative qualities, including nonlinear sampling strategies and agent-based modeling that offer enhancements over single-shot generations, whereby LLMs are used to both generate content and assess content against an objective target. Examples provided include complex question answering, code generation, and execution in the context of automated force-field development from actively learned density functional theory (DFT) modeling and data analysis.

PRefLexOR (Preference-based Recursive Language Modeling for Exploratory Optimization of Reasoning) combines preference optimization with concepts from Reinforcement Learning to enable models to self-teach through iterative reasoning improvements. We propose a recursive learning approach that engages the model in multi-step reasoning, revisiting, and refining intermediate steps before producing a final output in training and inference phases. Through multiple training stages, the model first learns to align its reasoning with accurate decision paths by optimizing the log odds between preferred and non-preferred responses. During this process, PRefLexOR builds a dynamic knowledge graph by generating questions from random text chunks and retrieval-augmentation to contextualize relevant details from the entire training corpus. In the second stage, preference optimization enhances model performance by using rejection sampling to fine-tune reasoning quality by continually producing in-situ training data while masking the reasoning steps. Recursive optimization within a thinking token framework introduces iterative feedback loops, where the model refines reasoning, achieving deeper coherence, consistency, and adaptability. Implemented in small language models with only 3 billion parameters, we should that even tiny models can iteratively teach themselves to reason with greater depth and reflectivity. Our implementation is straightforward and can be incorporated into any existing pretrained LLM. We focus our examples on applications in biological materials science and demonstrate the method in a variety of case studies that range from in-domain to cross-domain applications. Using reasoning strategies that include thinking and reflection modalities we build a multi-agent recursive self-improving inference approach to successively improve responses via repeated sampling in inference time.

In today's rapidly evolving landscape of Artificial Intelligence, large language models (LLMs) have emerged as a vibrant research topic. LLMs find applications in various fields and contribute significantly. Despite their powerful language capabilities, similar to pre-trained language models (PLMs), LLMs still face challenges in remembering events, incorporating new information, and addressing domain-specific issues or hallucinations. To overcome these limitations, researchers have proposed Retrieval-Augmented Generation (RAG) techniques, some others have proposed the integration of LLMs with Knowledge Graphs (KGs) to provide factual context, thereby improving performance and delivering more accurate feedback to user queries. Education plays a crucial role in human development and progress. With the technology transformation, traditional education is being replaced by digital or blended education. Therefore, educational data in the digital environment is increasing day by day. Data in higher education institutions are diverse, comprising various sources such as unstructured/structured text, relational databases, web/app-based API access, etc. Constructing a Knowledge Graph from these cross-data sources is not a simple task. This article proposes a method for automatically constructing a Knowledge Graph from multiple data sources and discusses some initial applications (experimental trials) of KG in conjunction with LLMs for question-answering tasks.

High performance is a crucial consideration when executing a complex analytic query on a massive semantic graph. In a semantic graph, vertices and edges carry attributes of various types. Analytic queries on semantic graphs typically depend on the values of these attributes; thus, the computation must view the graph through a filter that passes only those individual vertices and edges of interest. Knowledge Discovery Toolbox (KDT), a Python library for parallel graph computations, is customizable in two ways. First, the user can write custom graph algorithms by specifying operations between edges and vertices. These programmer-specified operations are called semiring operations due to KDT's underlying linear-algebraic abstractions. Second, the user can customize existing graph algorithms by writing filters that return true for those vertices and edges the user wants to retain during algorithm execution. For high productivity, both semiring operations and filters are written in a high-level language, resulting in relatively low performance due to the bottleneck of having to call into the Python virtual machine for each vertex and edge. In this work, we use the Selective Embedded JIT Specialization (SEJITS) approach to automatically translate semiring operations and filters defined by programmers into a lower-level efficiency language, bypassing the upcall into Python. We evaluate our approach by comparing it with the high-performance Combinatorial BLAS engine, and show our approach enables users to write in high-level languages and still obtain the high performance of low-level code. We also present a new roofline model for graph traversals, and show that our high-performance implementations do not significantly deviate from the roofline. Overall, we demonstrate the first known solution to the problem of obtaining high performance from a productivity language when applying graph algorithms selectively on semantic graphs.

Trends in novel manufacturing systems lead to an increased level of data availability and smart usage of these data. Nowadays, many approaches are available to use the data, but because of an increased flexibility of the systems the interaction between machines and humans has become a challenge. Humans have to browse through a huge amount of data, need knowledge about the machine and underlying algorithms to interpret the results; they cannot use their known terms for communication, we call it the conceptual gap. The user should be enabled to communicate with the machine on a more abstract level and in a more natural way. Therefore, a natural language layer is introduced to provide users with a familiar interaction interface. Underlying layers contain knowledge about the domain, the machines and how data can be accessed and processed. This enables users' questions such as “Are there any anomalies in the system?” to be answered. Answers are provided in natural language and evaluated with a test set of 204 questions.

External knowledge graphs (KGs) can be used to augment large language models (LLMs), while simultaneously providing an explainable knowledge base of facts that can be inspected by a human. This approach may be particularly valuable in domains where explainability is critical, like human trafficking data analysis. However, creating KGs can pose challenges. KGs parsed from documents may comprise explicit connections (those directly stated by a document) but miss implicit connections (those obvious to a human although not directly stated). To address these challenges, this preliminary research introduces the GAME-KG framework, standing for 'Gaming for Augmenting Metadata and Enhancing Knowledge Graphs.' GAME-KG is a federated approach to modifying explicit as well as implicit connections in KGs by using crowdsourced feedback collected through video games. GAME-KG is shown through two demonstrations: a Unity test scenario from Dark Shadows, a video game that collects feedback on KGs parsed from US Department of Justice (DOJ) Press Releases on human trafficking, and a following experiment where OpenAI's GPT-4 is prompted to answer questions based on a modified and unmodified KG. Initial results suggest that GAME-KG can be an effective framework for enhancing KGs while simultaneously providing an explainable set of structured facts verified by humans. © 2024 IEEE.

The life sciences have a strong need for carefully curated, semantically rich fact repositories. Knowledge harvesting from unstructured textual sources is currently performed by highly skilled curators who manually feed semantics into such databases as a result of deep understanding of the documents chosen to populate such repositories. As this is a slow and costly process, we here advocate an automatic approach to the generation of database contents which is based on JREX, a high performance relation extraction system. As a real-life example, we target REGULONDB, the world's largest manually curated reference database for the transcriptional regulation network of E. coli. We investigate in our study the performance of automatic knowledge capture from various literature sources, such as PUBMED abstracts and associated full text articles. Our results show that we can, indeed, automatically re-create a considerable portion of the REGULONDB database by processing the relevant literature sources. Hence, this approach might help curators widen the knowledge acquisition bottleneck in this field.

Knowledge Graphs (KGs) are large collections of structured data that can model real world knowledge and are important assets for the companies that employ them. KGs are usually constructed iteratively and often show a sparse structure. Also, as knowledge evolves, KGs must be updated and completed. Many automatic methods for KG Completion (KGC) have been proposed in the literature to reduce the costs associated with manual maintenance. Motivated by an industrial case study aiming to enrich a KG specifically designed for Natural Language Understanding tasks, this paper presents an overview of classical and modern deep learning completion methods. In particular, we delve into Large Language Models (LLMs), which are the most promising deep learning architectures. We show that their applications to KGC are affected by several shortcomings, namely they neglect the structure of KG and treat KGC as a classification problem. Such limitations, together with the brittleness of the LLMs themselves, stress the need to create KGC solutions at the interface between symbolic and neural approaches and lead to the way ahead for future research in intelligible corpus-based KGC. © 2023 CEUR-WS. All rights reserved.

In recent years, transformer-based models have emerged as powerful tools for natural language processing tasks, demonstrating remarkable performance in several domains. However, they still present significant limitations. These shortcomings become more noticeable when dealing with highly specific and complex concepts, particularly within the scientific domain. For example, transformer models have particular difficulties when processing scientific articles due to the domain-specific terminologies and sophisticated ideas often encountered in scientific literature. To overcome these challenges and further enhance the effectiveness of transformers in specific fields, researchers have turned their attention to the concept of knowledge injection. Knowledge injection is the process of incorporating outside knowledge into transformer models to improve their performance on certain tasks. In this paper, we present a comprehensive study of knowledge injection strategies for transformers within the scientific domain. Specifically, we provide a detailed overview and comparative assessment of four primary methodologies, evaluating their efficacy in the task of classifying scientific articles. For this purpose, we constructed a new benchmark including both 24K labelled papers and a knowledge graph of 9.2K triples describing pertinent research topics. We also developed a full codebase to easily re-implement all knowledge injection strategies in different domains. A formal evaluation indicates that the majority of the proposed knowledge injection methodologies significantly outperform the baseline established by Bidirectional Encoder Representations from Transformers.

In recent years, transformer-based models have emerged as powerful tools for natural language processing tasks, demonstrating remarkable performance in several domains. However, they still present significant limitations. These shortcomings become more noticeable when dealing with highly specific and complex concepts, particularly within the scientific domain. For example, transformer models have particular difficulties when processing scientific articles due to the domain-specific terminologies and sophisticated ideas often encountered in scientific literature. To overcome these challenges and further enhance the effectiveness of transformers in specific fields, researchers have turned their attention to the concept of knowledge injection. Knowledge injection is the process of incorporating outside knowledge into transformer models to improve their performance on certain tasks. In this paper, we present a comprehensive study of knowledge injection strategies for transformers within the scientific domain. Specifically, we provide a detailed overview and comparative assessment of four primary methodologies, evaluating their efficacy in the task of classifying scientific articles. For this purpose, we constructed a new benchmark including both 24K labelled papers and a knowledge graph of 9.2K triples describing pertinent research topics. We also developed a full codebase to easily re-implement all knowledge injection strategies in different domains. A formal evaluation indicates that the majority of the proposed knowledge injection methodologies significantly outperform the baseline established by Bidirectional Encoder Representations from Transformers. © 2024

Neural language models (LMs) have been extensively trained on vast corpora to store factual knowledge about various aspects of the world described in texts. Current technologies typically employ knowledge editing methods or specific prompts to modify LM outputs. However, existing knowledge editing methods are costly and inefficient, struggling to produce appropriate text. Additionally, prompt engineering is opaque and requires significant effort to find suitable prompts. To address these issues, we introduce a new method called PSPEM (Prefix Soft-Prompt Editing Method), that can be used for a lifetime with just one training. It resolves the inefficiencies and generalizability issues in knowledge editing methods and overcomes the opacity of prompt engineering by automatically seeking optimal soft prompts. Specifically, PSPEM adopts a prompt encoder and an encoding converter to compress and refine key information in prompts and adopts prompt alignment techniques to guide model generation, ensuring text consistency and adherence to the intended structure and content. We have validated the effectiveness of PSPEM through knowledge editing and attribute inserting. On the COUNTERFACT dataset, PSPEM achieved nearly 100% editing accuracy and demonstrated the highest level of fluency. We further analyzed the similarities between PSPEM and original prompts and their impact on the model's internals. The results indicate that PSPEM can serve as an alternative to original prompts, supporting the model in effective editing.

Masked language models have quickly become the de facto standard when processing text. Recently, several approaches have been proposed to further enrich word representations with external knowledge sources such as knowledge graphs. However, these models are devised and evaluated in a monolingual setting only. In this work, we propose a language-independent entity prediction task as an intermediate training procedure to ground word representations on entity semantics and bridge the gap across different languages by means of a shared vocabulary of entities. We show that our approach effectively injects new lexical-semantic knowledge into neural models, improving their performance on different semantic tasks in the zero-shot crosslingual setting. As an additional advantage, our intermediate training does not require any supplementary input, allowing our models to be applied to new datasets right away. In our experiments, we use Wikipedia articles in up to 100 languages and already observe consistent gains compared to strong baselines when predicting entities using only the English Wikipedia. Further adding extra languages lead to improvements in most tasks up to a certain point, but overall we found it non-trivial to scale improvements in model transferability by training on ever increasing amounts of Wikipedia languages. © 2021 Association for Computational Linguistics.

We present a method, which incorporates knowledge awareness into the symbolic computation of discrete controllers for reactive cyber physical systems, to improve decision making about the unknown operating environment under uncertain/incomplete inputs. Assuming an abstract model of the system and the environment, we translate the knowledge awareness of the operating context into linear temporal logic formulas and incorporate them into the system specifications to synthesize a controller. The knowledge base is built upon an ontology model of the environment objects and behavioural rules, which includes also symbolic models of partial input features. The resulting symbolic controller support smoother, early reactions, which improves the security of the system over existing approaches based on incremental symbolic perception. A motion planning case study for an autonomous vehicle has been implemented to validate the approach, and presented results show significant improvements with respect to safety of state-of-the-art symbolic controllers for reactive systems.

Memory is one of the most essential cognitive functions serving as a repository of world knowledge and episodes of activities. In recent years, large-scale pre-trained language models have shown remarkable memorizing ability. On the contrary, vanilla neural networks without pre-training have been long observed suffering from the catastrophic forgetting problem. To investigate such a retentive-forgetful contradiction and understand the memorizing dynamic mechanism of language models, we conduct thorough experiments by controlling the target knowledge types, the learning strategies and the learning schedules. We find that: 1) Vanilla language models without pre-training are forgetful; 2) Pre-training leads to retentive language models; 3) Knowledge relevance and diversification significantly influence the memory formation. These conclusions are useful for understanding the abilities of pre-trained language models and shed light on designing and evaluating new learning and inference algorithms of language models. © 2024 ELRA Language Resource Association: CC BY-NC 4.0.

Previous literatures show that pre-trained masked language models (MLMs) such as BERT can achieve competitive factual knowledge extraction performance on some datasets, indicating that MLMs can potentially be a reliable knowledge source. In this paper, we conduct a rigorous study to explore the underlying predicting mechanisms of MLMs over different extraction paradigms. By investigating the behaviors of MLMs, we find that previous decent performance mainly owes to the biased prompts which overfit dataset artifacts. Furthermore, incorporating illustrative cases and external contexts improve knowledge prediction mainly due to entity type guidance and golden answer leakage. Our findings shed light on the underlying predicting mechanisms of MLMs, and strongly question the previous conclusion that current MLMs can potentially serve as reliable factual knowledge bases(1).

Rare diseases affect millions worldwide but often face limited research focus due to their low prevalence. This results in prolonged diagnoses and a lack of approved therapies. Recent advancements in Large Language Models (LLMs) have shown promise in automating the extraction of medical information, offering potential to improve medical diagnosis and management. However, most LLMs lack professional medical knowledge, especially concerning rare diseases, and struggle to handle the latest rare disease information. They also cannot effectively manage rare disease data and are not directly suitable for diagnosis and management tasks. Our objective is to create an end-to-end system called AutoRD, which automates the extraction of information from medical texts about rare diseases, focusing on entities and their relations. AutoRD integrates up-to-date structured knowledge and demonstrates superior performance in rare disease extraction tasks. We conduct various experiments to evaluate AutoRD's performance, aiming to surpass common LLMs and traditional methods.

The efficient management and utilization of coal mine equipment maintenance knowledge is an indispensable foundation for advancing the establishment of intelligent mines. This knowledge has problems such as scattered, low sharing, and insufficient management, which restricts the development of coal mine intelligence. For the above-mentioned problems, a large language model for the maintenance of coal mine equipment based on multi-source text (XCoalChat) was proposed to better manage and utilize the existing massive knowledge of coal mine equipment maintenance. The dataset of coal mine equipment maintenance based on ReliableCEMK-Self-Instruction was constructed to obtain a wide and diverse amount of knowledge through sample generation. Aiming at the illusory problem of the large language model, a knowledge graph enhancement method based on the "Coal Mine Equipment Maintenance System-Full Life Cycle-Specification" was proposed to improve the knowledge density. A triple-LoRA fine-tuning mechanism and DPO direct preference optimization method were introduced into the top of the baseline model, which guarantees that XCoalChat can handle multiple Q&A and maintenance decision analysis tasks with limited computing power. Compared with ChatGLM, Bloom, and LLama, the comprehensive assessment of XCoalChat was performed by experiments including coal mine dialog consulting, coal mine professional consulting, and maintenance decision analysis. The results showed that XCoalChat achieved the best response accuracy in professional consulting and maintenance decision analysis; XCoalChat also took the least reasoning time on average. XCoalChat outperformed other mainstream large language models, which verify that XCoalChat is an effective large language model in the field of coal mine equipment maintenance.

In the current Digital Transformation scenario, Knowledge Graphs are essential for comprehending, representing, and exploiting complex information in a structured form. The main paradigm in automatically generating proper Knowledge Graphs relies on predefined schemas or ontologies. Such schemas are typically manually constructed, requiring an intensive human effort, and are often sensitive to information loss due to negligence, incomplete analysis, or human subjectivity or inclination. Limiting human bias and the resulting information loss in creating proper Knowledge Graphs is paramount, particularly for user modeling in various sectors, such as education or healthcare. To this end, we propose a novel approach to automatically generating a proper entity schema. The devised methodology combines the language understanding capabilities of LLM with classical machine learning methods such as clustering to properly build an entity schema from a set of documents. This solution eliminates the need for human intervention and fosters a more efficient and comprehensive knowledge representation. The assessment of our proposal concerns adopting a state-of-the-art entity extraction model ( UniNER) to estimate the relevance of the extracted entities based on the generated schema. Results confirm the potential of our approach, as we observed a negligible difference between the topic similarity score obtained with the ground truth and with the automatically generated schema (less than 1% on average on three different datasets). Such an outcome confirms that the proposed approach may be valuable in automatically creating an entity schema from a set of documents.

In the current digitalization era, capturing and effectively representing knowledge is crucial in most real-world scenarios. In this context, knowledge graphs represent a potent tool for retrieving and organizing a vast amount of information in a properly interconnected and interpretable structure. However, their generation is still challenging and often requires considerable human effort and domain expertise, hampering the scalability and flexibility across different application fields. This paper proposes an innovative knowledge graph generation approach that leverages the potential of the latest generative large language models, such as GPT-3.5, that can address all the main critical issues in knowledge graph building. The approach is conveyed in a pipeline that comprises novel iterative zero-shot and external knowledge-agnostic strategies in the main stages of the generation process. Our unique manifold approach may encompass significant benefits to the scientific community. In particular, the main contribution can be summarized by: (i) an innovative strategy for iteratively prompting large language models to extract relevant components of the final graph; (ii) a zero-shot strategy for each prompt, meaning that there is no need for providing examples for "guiding" the prompt result; (iii) a scalable solution, as the adoption of LLMs avoids the need for any external resources or human expertise. To assess the effectiveness of our proposed model, we performed experiments on a dataset that covered a specific domain. We claim that our proposal is a suitable solution for scalable and versatile knowledge graph construction and may be applied to different and novel contexts.

SNOMED CT postcoordination is an underused mechanism that can help to implement advanced systems for the automatic extraction and encoding of clinical information from text. It allows defining non-existing SNOMED CT concepts by their relationships with existing ones. Manually building postcoordinated expressions is a difficult task. It requires a deep knowledge of the terminology and the support of specialized tools that barely exist. In order to support the building of postcoordinated expressions, we have implemented KGE4SCT: a method that suggests the corresponding SNOMED CT postcoordinated expression for a given clinical term. We leverage on the SNOMED CT ontology and its graph-like structure and use knowledge graph embeddings (KGEs). The objective of such embeddings is to represent in a vector space knowledge graph components (e.g. entities and relations) in a way that captures the structure of the graph. Then, we use vector similarity and analogies for obtaining the postcoordinated expression of a given clinical term. We obtained a semantic type accuracy of 98%, relationship accuracy of 90%, and analogy accuracy of 60%, with an overall completeness of postcoordination of 52% for the Spanish SNOMED CT version. We have also applied it to the English SNOMED CT version and outperformed state of the art methods in both, corpus generation for language model training for this task (improvement of 6% for analogy accuracy), and automatic postcoordination of SNOMED CT expressions, with an increase of 17% for partial conversion rate.

Motivation Creating knowledge bases and ontologies is a time consuming task that relies on manual curation. AI/NLP approaches can assist expert curators in populating these knowledge bases, but current approaches rely on extensive training data, and are not able to populate arbitrarily complex nested knowledge schemas.Results Here we present Structured Prompt Interrogation and Recursive Extraction of Semantics (SPIRES), a Knowledge Extraction approach that relies on the ability of Large Language Models (LLMs) to perform zero-shot learning and general-purpose query answering from flexible prompts and return information conforming to a specified schema. Given a detailed, user-defined knowledge schema and an input text, SPIRES recursively performs prompt interrogation against an LLM to obtain a set of responses matching the provided schema. SPIRES uses existing ontologies and vocabularies to provide identifiers for matched elements. We present examples of applying SPIRES in different domains, including extraction of food recipes, multi-species cellular signaling pathways, disease treatments, multi-step drug mechanisms, and chemical to disease relationships. Current SPIRES accuracy is comparable to the mid-range of existing Relation Extraction methods, but greatly surpasses an LLM's native capability of grounding entities with unique identifiers. SPIRES has the advantage of easy customization, flexibility, and, crucially, the ability to perform new tasks in the absence of any new training data. This method supports a general strategy of leveraging the language interpreting capabilities of LLMs to assemble knowledge bases, assisting manual knowledge curation and acquisition while supporting validation with publicly-available databases and ontologies external to the LLM.Availability and implementation SPIRES is available as part of the open source OntoGPT package: https://github.com/monarch-initiative/ontogpt.

Exploration and self-observation are key mechanisms of infant sensorimotor development. These processes are further guided by parental scaffolding to accelerate skill and knowledge acquisition. In developmental robotics, this approach has been adopted often by having a human acting as the source of scaffolding. In this study, we investigate whether Large Language Models (LLMs) can act as a scaffolding agent for a robotic system that aims to learn to predict the effects of its actions. To this end, an object manipulation setup is considered where one object can be picked and placed on top of or in the vicinity of another object. The adopted LLM is asked to guide the action selection process through algorithmically generated state descriptions and action selection alternatives in natural language. The simulation experiments that include cubes in this setup show that LLM-guided (GPT3.5-guided) learning yields significantly faster discovery of novel structures compared to random exploration. However, we observed that GPT3.5 fails to effectively guide the robot in generating structures with different affordances such as cubes and spheres. Overall, we conclude that even without fine-tuning, LLMs may serve as a moderate scaffolding agent for improving robot learning, however, they still lack affordance understanding which limits the applicability of the current LLMs in robotic scaffolding tasks.

Knowledge about the interactions between dietary and biomedical factors is scattered throughout uncountable research articles in an unstructured form (e.g., text, images, etc.) and requires automatic structuring so that it can be provided to medical professionals in a suitable format. Various biomedical knowledge graphs exist, however, they require further extension with relations between food and biomedical entities. In this study, we evaluate the performance of three state-of-the-art relation-mining pipelines (FooDis, FoodChem and ChemDis) which extract relations between food, chemical and disease entities from textual data. We perform two case studies, where relations were automatically extracted by the pipelines and validated by domain experts. The results show that the pipelines can extract relations with an average precision around 70%, making new discoveries available to domain experts with reduced human effort, since the domain experts should only evaluate the results, instead of finding, and reading all new scientific papers.

Augmenting pre-trained language models with knowledge graphs (KGs) has achieved success on various commonsense reasoning tasks. However, for a given task instance, the KG, or certain parts of the KG, may not be useful. Although KG-augmented models often use attention to focus on specific KG components, the KG is still always used, and the attention mechanism is never explicitly taught which KG components should be used. Meanwhile, saliency methods can measure how much a KG feature (e.g., graph, node, path) influences the model to make the correct prediction, thus explaining which KG features are useful. This paper explores how saliency explanations can be used to improve KG-augmented models' performance. First, we propose to create coarse (Is the KG useful?) and fine (Which nodes/paths in the KG are useful?) saliency explanations. Second, to motivate saliency-based supervision, we analyze oracle KG-augmented models which directly use saliency explanations as extra inputs for guiding their attention. Third, we propose SALKG, a framework for KG-augmented models to learn from coarse and/or fine saliency explanations. Given saliency explanations created from a task's training set, SALKG jointly trains the model to predict the explanations, then solve the task by attending to KG features highlighted by the predicted explanations. On three commonsense QA benchmarks (CSQA, OBQA, CODAH) and a range of KG-augmented models, we show that SALKG can yield considerable performance gains - up to 2.76% absolute improvement on CSQA. (2)